Re: [PATCH v1 4/5] iommu/arm-smmu-v3: Keep track of attached ssids

[Jason Gunthorpe <jgg@nvidia.com>]

On Thu, May 11, 2023 at 04:50:51AM +0800, Michael Shavit wrote:

> @@ -213,14 +213,14 @@ static void arm_smmu_mm_invalidate_range(struct mmu_notifier *mn,
>  	if (!(smmu_domain->smmu->features & ARM_SMMU_FEAT_BTM))
>  		arm_smmu_tlb_inv_range_asid(start, size, smmu_mn->cd->asid,
>  					    PAGE_SIZE, false, smmu_domain);
> -	arm_smmu_atc_inv_domain(smmu_domain, mm->pasid, start, size);
> +	arm_smmu_atc_inv_domain_ssid(smmu_domain, mm->pasid, start,
>  	size);

You should be getting rid of mm->pasid in this series as well.

When each domain keeps track of what STE/CD entries that point to it then
*ALL* invalidation should iterate over the list of pointing entires
and generate the correct invalidation for that pointer.

Eg we learn the PASID from the fact that a CD at PASID xyz is pointing
at this domain and generate an invalidation for that PASID.

mm->pasid is logically incorrect in all of this code with our
multi-attach model, it was here because this code wasn't tracking at
what as pointing at the iommu_domain.

Jason